Benchmark Almanac

CNFinBench

Doubao Seed 2.1 Pro 83.7 leads since June 2026
Not in model cards

Models answer financial questions and carry out financial workflows. The benchmark includes tasks on expertise, autonomy, and integrity.

Released
December 2025
Kind
Finance
Level
Expert
Scored
Settled by events
Answers
Agent task
Tasks
All public
Measures
Capability
Language
Chinese
In and out
Text → Text, Actions
Interaction
Agentic
Environment
APIs and tools
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude Opus 5.5 #5 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
13
Entries on its own leaderboard, 10 models we track

Doubao Seed 2.1 Pro leads with 83.7 among models released by September 2026 Doubao Seed 2.1 Pro leads with 83.7 among models released by September 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Doubao Seed 2.1 Pro LLM 83.7
  2. 2 Kimi K3 LLM 82.9
  3. 3 DeepSeek V4.1 Flash LLM 80.6
  4. 4 GPT-6.1 Sol LLM 80.6
  5. 5 Intern-S2-Preview-397B LLM 80.3
  6. 6 Gemini 3.1 Pro LLM 80.0
  7. 7 Claude Opus 5.5 LLM 79.1
  8. 8 GPT-5.6 Sol Agent 78.1
  9. 9 MTFin-Qwen3.8-V2 LLM 76.0
  10. 10 Qwen 3.8 Max LLM 74.9
  11. 11 MiniMax-M3 LLM 69.5
  12. 12 Dowsure-Fin-27B-V1 LLM 60.2
  13. 13 DeepSeek-V4-Pro Agent 52.7

Source: Official CNFinBench leaderboard

#EntryScoreSettingSubmitted byReleasedRun onSource
1 Doubao Seed 2.1 Pro 83.7 LLM ByteDance Jun 2026 Oct 2026 Official leaderboard
2 Kimi K3 82.9 LLM Moonshot Jul 2026 Oct 2026 Official leaderboard
3 DeepSeek V4.1 Flash 80.6 LLM DeepSeek AI Sep 2026 Oct 2026 Official leaderboard
4 GPT-6.1 Sol 80.6 LLM OpenAI Sep 2026 Oct 2026 Official leaderboard
5 Intern-S2-Preview-397B open weights, 397B 80.3 LLM Shanghai AI Laboratory Oct 2026 Official leaderboard
6 Gemini 3.1 Pro 80.0 LLM Google Feb 2026 Oct 2026 Official leaderboard
7 Claude Opus 5.5 79.1 LLM Anthropic Sep 2026 Oct 2026 Official leaderboard
8 GPT-5.6 Sol 78.1 Agent OpenAI Jul 2026 Oct 2026 Official leaderboard
9 MTFin-Qwen3.8-V2 closed weights, 27B 76.0 LLM 美团金融大模型应用中心 Nov 2026 Official leaderboard
10 Qwen 3.8 Max 74.9 LLM Alibaba Aug 2026 Oct 2026 Official leaderboard
11 MiniMax-M3 69.5 LLM MiniMax Jun 2026 Oct 2026 Official leaderboard
12 Dowsure-Fin-27B-V1 closed weights, 27B 60.2 LLM Dowsure Nov 2026 Official leaderboard
13 DeepSeek-V4-Pro 52.7 Agent DeepSeek Apr 2026 Oct 2026 Official leaderboard
#ModelScoreSettingSubmitted byReleasedRun onSource
1 Doubao Seed 2.1 Pro 83.7 LLM ByteDance Jun 2026 Oct 2026 Official leaderboard
2 Kimi K3 82.9 LLM Moonshot Jul 2026 Oct 2026 Official leaderboard
3 DeepSeek V4.1 Flash 80.6 LLM DeepSeek AI Sep 2026 Oct 2026 Official leaderboard
4 GPT-6.1 Sol 80.6 LLM OpenAI Sep 2026 Oct 2026 Official leaderboard
5 Gemini 3.1 Pro 80.0 LLM Google Feb 2026 Oct 2026 Official leaderboard
6 Claude Opus 5.5 79.1 LLM Anthropic Sep 2026 Oct 2026 Official leaderboard
7 GPT-5.6 Sol 78.1 Agent OpenAI Jul 2026 Oct 2026 Official leaderboard
8 Qwen 3.8 Max 74.9 LLM Alibaba Aug 2026 Oct 2026 Official leaderboard
9 MiniMax-M3 69.5 LLM MiniMax Jun 2026 Oct 2026 Official leaderboard
10 DeepSeek-V4-Pro 52.7 Agent DeepSeek Apr 2026 Oct 2026 Official leaderboard

Rankings by part

We read 21 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as 金融专业知识问答, Fin_Basics (Accuracy) and Fin_Cert_Exams (Accuracy). Doubao Seed 2.1 Pro leads 8 of them, and 5 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

金融专业知识问答

8 models we track have a score here.

  1. 1 Doubao Seed 2.1 Pro 92.7
  2. 2 Kimi K3 91.4
  3. 3 Gemini 3.1 Pro ↑2 90.3
  4. 4 Claude Opus 5.5 ↑2 89.7
  5. 5 GPT-6.1 Sol ↓1 89.0
  6. 6 Qwen 3.8 Max ↑2 88.0
  7. 7 DeepSeek V4.1 Flash ↓4 88.0
  8. 8 MiniMax-M3 ↑1 71.0

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Doubao Seed 2.1 Pro
Agent LLM
83.7
Kimi K3
Agent LLM
82.9
Gemini 3.1 Pro
Agent LLM
80.0
Qwen 3.8 Max
Agent LLM
74.9

Source: Official CNFinBench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Doubao Seed 2.1 Pro LLM 83.7
Kimi K3 LLM 82.9
DeepSeek V4.1 Flash LLM 80.6
GPT-6.1 Sol LLM 80.6
Gemini 3.1 Pro LLM 80.0
Claude Opus 5.5 LLM 79.1
GPT-5.6 Sol Agent 78.1
Qwen 3.8 Max LLM 74.9
MiniMax-M3 LLM 69.5
Kimi K3 Agent 69.3
Qwen 3.8 Max Agent 63.0
Doubao Seed 2.1 Pro Agent 60.2
Gemini 3.1 Pro Agent 54.3
DeepSeek-V4-Pro Agent 52.7

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Doubao Seed 2.1 Pro
Agent LLM
83.7
Kimi K3
Agent LLM
82.9
Gemini 3.1 Pro
Agent LLM
80.0
Qwen 3.8 Max
Agent LLM
74.9

Source: Official CNFinBench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Doubao Seed 2.1 Pro LLM 83.7
Kimi K3 LLM 82.9
DeepSeek V4.1 Flash LLM 80.6
GPT-6.1 Sol LLM 80.6
Gemini 3.1 Pro LLM 80.0
Claude Opus 5.5 LLM 79.1
GPT-5.6 Sol Agent 78.1
Qwen 3.8 Max LLM 74.9
MiniMax-M3 LLM 69.5
Kimi K3 Agent 69.3
Qwen 3.8 Max Agent 63.0
Doubao Seed 2.1 Pro Agent 60.2
Gemini 3.1 Pro Agent 54.3
DeepSeek-V4-Pro Agent 52.7

2 models have held number one on the official leaderboard since its release in December 2025

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind ByteDance Seed
Own leaderboard Not tested since Sep 2026

Source: Official CNFinBench leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
50.0 60.0 70.0 80.0 90.0 2026 50.0 60.0 70.0 80.0 90.0 2026

Source: Official CNFinBench leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (13 entries).
  • Pass Models scored 10 models we track, plus 3 other entries on its leaderboard.
  • Pass Newest score The newest is from September 2026, for GPT-6.1 Sol.
  • Note Frontier models Only 2 of the top 10 have a score. The strongest tested is Claude Opus 5.5 (#5).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on cnfinbench.opencompass.org.cn: 14 rows for models we track and 3 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check